-
HistoKT: Cross Knowledge Transfer in Computational Pathology
Authors:
Ryan Zhang,
Jiadai Zhu,
Stephen Yang,
Mahdi S. Hosseini,
Angelo Genovese,
Lina Chen,
Corwyn Rowsell,
Savvas Damaskinos,
Sonal Varma,
Konstantinos N. Plataniotis
Abstract:
The lack of well-annotated datasets in computational pathology (CPath) obstructs the application of deep learning techniques for classifying medical images. %Since pathologist time is expensive, dataset curation is intrinsically difficult. Many CPath workflows involve transferring learned knowledge between various image domains through transfer learning. Currently, most transfer learning research…
▽ More
The lack of well-annotated datasets in computational pathology (CPath) obstructs the application of deep learning techniques for classifying medical images. %Since pathologist time is expensive, dataset curation is intrinsically difficult. Many CPath workflows involve transferring learned knowledge between various image domains through transfer learning. Currently, most transfer learning research follows a model-centric approach, tuning network parameters to improve transfer results over few datasets. In this paper, we take a data-centric approach to the transfer learning problem and examine the existence of generalizable knowledge between histopathological datasets. First, we create a standardization workflow for aggregating existing histopathological data. We then measure inter-domain knowledge by training ResNet18 models across multiple histopathological datasets, and cross-transferring between them to determine the quantity and quality of innate shared knowledge. Additionally, we use weight distillation to share knowledge between models without additional training. We find that hard to learn, multi-class datasets benefit most from pretraining, and a two stage learning framework incorporating a large source domain such as ImageNet allows for better utilization of smaller datasets. Furthermore, we find that weight distillation enables models trained on purely histopathological features to outperform models using external natural image data.
△ Less
Submitted 26 January, 2022;
originally announced January 2022.
-
End-to-End Models for the Analysis of System 1 and System 2 Interactions based on Eye-Tracking Data
Authors:
Alessandro Rossi,
Sara Ermini,
Dario Bernabini,
Dario Zanca,
Marino Todisco,
Alessandro Genovese,
Antonio Rizzo
Abstract:
While theories postulating a dual cognitive system take hold, quantitative confirmations are still needed to understand and identify interactions between the two systems or conflict events. Eye movements are among the most direct markers of the individual attentive load and may serve as an important proxy of information. In this work we propose a computational method, within a modified visual vers…
▽ More
While theories postulating a dual cognitive system take hold, quantitative confirmations are still needed to understand and identify interactions between the two systems or conflict events. Eye movements are among the most direct markers of the individual attentive load and may serve as an important proxy of information. In this work we propose a computational method, within a modified visual version of the well-known Stroop test, for the identification of different tasks and potential conflicts events between the two systems through the collection and processing of data related to eye movements. A statistical analysis shows that the selected variables can characterize the variation of attentive load within different scenarios. Moreover, we show that Machine Learning techniques allow to distinguish between different tasks with a good classification accuracy and to investigate more in depth the gaze dynamics.
△ Less
Submitted 3 February, 2020;
originally announced February 2020.
-
The Andromeda Study: A Femto-Spacecraft Mission to Alpha Centauri
Authors:
Andreas M. Hein,
Kelvin F. Long,
Dan Fries,
Nikolaos Perakis,
Angelo Genovese,
Stefan Zeidler,
Martin Langer,
Richard Osborne,
Rob Swinney,
John Davies,
Bill Cress,
Marc Casson,
Adrian Mann,
Rachel Armstrong
Abstract:
This paper discusses the physics, engineering and mission architecture relating to a gram-sized interstellar probe propelled by a laser beam. The objectives are to design a fly-by mission to Alpha Centauri with a total mission duration of 50 years travelling at a cruise speed of 0.1c. Furthermore, optical data from the target star system is to be obtained and sent back to the Solar system. The mai…
▽ More
This paper discusses the physics, engineering and mission architecture relating to a gram-sized interstellar probe propelled by a laser beam. The objectives are to design a fly-by mission to Alpha Centauri with a total mission duration of 50 years travelling at a cruise speed of 0.1c. Furthermore, optical data from the target star system is to be obtained and sent back to the Solar system. The main challenges of such a mission are presented and possible solutions proposed. The results show that by extrapolating from currently existing technology, such a mission would be feasible. The total mass of the proposed spacecraft is 23g and the space-based laser infrastructure has a beam power output of 15GW. Rurther exploration of the laser - spacecraft tradespace and associated technologies are necessary.
△ Less
Submitted 11 August, 2017;
originally announced August 2017.
-
Experimental Evidences for Static Charge Density Waves in Iron Oxy-pnictides
Authors:
A. Martinelli,
P. Manfrinetti,
A. Provino,
A. Genovese,
F. Caglieris,
G. Lamura,
C. Ritter,
M. Putti
Abstract:
In this Letter we report high-resolution synchrotron X-ray powder diffraction and transmission electron microscope analysis of Mn-substituted LaFeAsO samples, demonstrating that a static incommensurate modulated structure develops across the low-temperature orthorhombic phase, whose modulation wave-vector depends on the Mn content. The incommensurate structural distortion is likely originating fro…
▽ More
In this Letter we report high-resolution synchrotron X-ray powder diffraction and transmission electron microscope analysis of Mn-substituted LaFeAsO samples, demonstrating that a static incommensurate modulated structure develops across the low-temperature orthorhombic phase, whose modulation wave-vector depends on the Mn content. The incommensurate structural distortion is likely originating from a charge-density-wave instability, a periodic modulation of the density of conduction electrons associated with a modulation of the atomic positions. Our results add a new component in the physics of Fe-based superconductors, indicating that the density wave ordering is charge-driven.
△ Less
Submitted 13 December, 2016;
originally announced December 2016.
-
Role of heat and mechanical treatments in the fabrication of superconducting Ba0.6K0.4Fe2As2 ex-situ Powder-In-Tube tapes
Authors:
A. Malagoli,
E. Wiesenmayer,
S. Marchner,
D. Johrendt,
A. Genovese,
M. Putti
Abstract:
Among the recently discovered Fe-based superconducting compounds, the (K,Ba)Fe2As2 phase is attracting large interest within the scientific community interested in conductor developments. In fact, after some years of development, critical current densities Jc of about 105 A/cm2 at fields up to more than 10 T have been obtained in powder in tube (PIT) processed wires and tapes. Here we explore the…
▽ More
Among the recently discovered Fe-based superconducting compounds, the (K,Ba)Fe2As2 phase is attracting large interest within the scientific community interested in conductor developments. In fact, after some years of development, critical current densities Jc of about 105 A/cm2 at fields up to more than 10 T have been obtained in powder in tube (PIT) processed wires and tapes. Here we explore the crucial points in the wire/tape fabrication by means of the ex-situ PIT method. We focus on scaling up processes which are crucial for the industrial fabrication. We analyzed the effects on the microstructure of the different heat and mechanical treatments. By an extensive microstructural analysis correlated with the transport properties we addressed the issues concerning the phase purity, the internal porosity and crack formation in the superconducting core region. Our best conductors with a filling factor of about 30 heat treated at 800 C exhibited Tc = 38 K the highest value measured in such kind of superconducting tape. The microstructure analysis shows clean and well connected grain boundaries but rather poor density: The measured Jc of about 3 x 10^4 A/cm2 in self-field is suppressed by less than a factor 7 at 7 T. Such not yet optimized Jc values can be accounted for by the reduced density while the moderate in-field suppression and a rather high n-factor confirm the high homogeneity and uniformity of these tapes.
△ Less
Submitted 14 July, 2015;
originally announced July 2015.
